Inspiration

Remember that day when you are in a supermarket, you have a shopping list on your hand, and you know exactly what you need, but you got to go find it? Now, imagine our shopper helper in every isle. Instead of looking through all the price tag to find the item that meets your need, you can just say a sentence and our shopper helper would just do all the heavy lifting for you - you could just say: "Hey, what's the cheapest cereal in the store?". And, shopper helper would reply: "It's the Oompa Loompa Cereal and it is $2.10" in a very intuitive conversation.

What it does

First, Shopper Helper takes in customer voice as an input and translate it into text. Second, Shopper Helper saves customer time. It allows employees to focus on their main task. And, most importantly, increase user experience and satisfaction.

How we built it

First, we implement the bing speech to text API to capture the customer request. This request is then sent to Microsoft LUIS AI analysis. LUIS returns the intent and entities, which will aid in querying a request in NCR. We then make request to NCR (end points such as items/prices and suggestions to name a few) to obtain the answer to the user's request. After some comparisons and analysis in python, the results is then converted to a string and lastly converted to audio, which is then played back the user. All backend was done through python and requests module.

Accomplishments that we're proud of

Getting bing speech to text to work with high accuracy Train Luis AI to classify over 10 intents Troubleshooting obstacles and find a path to successfully integrate sponsors API Most importantly, we are proud that we had created a product that alleviates customer frustration, employee workload, and modernize shopping experience.

What's next for ShopperHelper 5000

Test out the product in small and medium sized enterprises - following lean methodologies - built, measure, and learn.

Built With

Share this project:
×

Updates